8/2/2019 01394737

1/7

In this article we provide an understanding of M2M, !i&Z: _ rdiscuss the fundamental issues, and propose a strategic

-approach f arly development to successful deploymentof a resilient M2M solution.

WHY ADOPT MZM?

..

40 IEE Computing & Control Engineering IOaoberINovember 2004

8/2/2019 01394737

2/7

Common problems that arise in manual data collectionsystems include downtime issues which impact customer

service, slow response to field problems, operational andsupply chain inefficiencies, ost revenues and many more,depending on the type of machine and who owns it.

These problems can be solved by early access toinformation that is perhaps already contained within themachine. These data may i d e n t i for example, out-of-bandoperation; failure conditions; equipment use; maintenanceand service n eedd stat us; on-time; consumables status;user/operator ID: nd configuration and set-up.

Furthermore, M2 M connectivity deployed to solve aspecific problem enables benefits to be derived from theadditional information obtained. For example, data iivm amachine could be used by OEM manufacturers to hpr0Veproduct reliability, or consumable suppliers to improvesupply chain management. Service operators could use thedata to minimise machine downtime.

This presents new ways of extending companies'businesses based on the value of the data collected. n manycases, t enahl s a fundamental change in the bu shes modelbased on service rather than the traditional sale ofmanufactu red goods. With timely and co rrect data, forexample, a company might migrate to a leased saleintegrating service and maintenance costs or pay per usagewhether costed on a time basis or on other machine sWifIc

*arametem such as revolutions of a turbine'.

IEE Computing 8 Control Engineering IOboberINawmber 2004 41

8/2/2019 01394737

3/7

WHAT IS M2M?Machine-to-machine is the automated communicationbetween remote machines and central managementapplications. It provides real-time control and monitoringwithout human intervention, by removing the manualp m s s nd the associated 'paper trail'.

As a result, operations are more efficient, which yieldscost savings and the increased levels of customer serviceaids in customer retention. Revenues can be i n m a d basedm u n d service, where incremeutal seMces m nabled anddelivered over the product's lifetime.

The applications of M7M are numerous. M2M is suitedto any machine which can be controlled or where stat usinformation can be obtained or externally sensed. It couldbe as simple as a light switch through to a complexindustrial machine. Typical examples include HVAC,electrical systems, industr ial automation, utility, officebusiness systems, telematics, lighting, security, vending,gaming, and domestic appliances to name a few.

Conversely, many cent ral management applicationsalready exist. They include supply chain management(SCM), enterpris e resource planning (ERP), customerrelationship management (CRM),and billing and paymentsystems. Although new markets enable web-basedapplications to be specifically created, exist ing legacy ITsystems can be extended to integrate the machine level datadirectly

There are many elements within M2M (see table'Elements within 1M2M', 43) o address the 'end-bend ' needbetween remote machines and central applications.

PITFALLS AN D FUNDAMENTAL ISSUESM2M technology leverages the existing infrastructure toprovide the core communication path between remotemachines and central management applications. However

there is a problem with the existing infrastrudure. Wreless

mobile networks were largely des ignd for voice tratfic, andwireless mobile GPRS data networks were designed formobile 'client' data user s such as PDAs. Wreless terminalproviders leverage mobile telephone technology TheInternet, ISPs and IT departments deliver IT services andsolutions for PC and human users.

In short, the Communications infrastructure wasdesigned for human users, who have intelligence (usually!)and c an make decisions. The resilience of the entirei n f " c t u r e is underpinned by the human ability to makesimple decisions to recover from adverse conditions, likere-booti the PC.

Canverselx the vast majorily of machines is not designedfor remote connectivity Where they a re connected' theydeliver small amounts of data on an insrepuent basis and inmany cases remain fmed to the Same location. Under thesecircumstances, the GSWGPRS network may dengist er themachine rendering it 'deaf' to incoming calls, where itthinks that the machine is really a mobile phone tha t hasmoved into another cell.

There are many issues to consider. Traditionalcommunications methods rely on central pollingtechniques, however thi s ha s the drawback of creating a'closed' system based on proprietary solutions that do notinteroperate with thi rd parties. This typically requires aprivate infrastructure and results in central sitebottlenecks as the application sequentially processes eachremote device.

Also, the Internet is client driven, which means tha t itrepuires he remote machine (or M7.M hardware) to initiatethe connection to the central application. ISPs do nottraditionally initiate calls due to the incurred call harges.

And, although the GPRS network is based on TCP/P, itis largely a private network with a fm wdl present betweenthe private GPRS network and the Internet. As a result, the

remote machine is not addressable from the Internet

11 I I EComputing 8 control Engineering IOcfoberINavember2W 4

8/2/2019 01394737

4/7

M2M

M l M hardware

ELEMENTS WITHIN M2M

This hardware adapts the rem ote m achine (through the varied interfaces) to th e netw ork,typically using mod em technology.It can act as a gateway to the network where numerousmachines are locally networked behind the M2M hardware. Depending on the level ofintelligence with in the remote m achine, the M 2M ha rdware may contain TCP/IP, datatrans port p rotoc ols such as HTTP, FTP, SMTP and a pplica tion layer 'Device M ana gem enfprotocols for transporting data directly intoa central application.

Element

Remote machine

Central applications

User devices

Description

This can be any type of machine that can be electronically connected to a commu nicationsline. The various interface s can range from externally sensed I/O through to intelligentproto col based interfaces, wh ich can be proprietary, ma rket specific, or 'open standard', suchas TCP/IP.

This presents meaningful inform ation to users about the machine ne twork as a whole.Information at any levelis provided on-dem and or through automa ted services. Data may beintegrated directly with in e nterpriseIT systems such as ERP by middlew are or offered as astandalone application or service.

Analysed inform ation is available online or through automa ted rep orts to user devices suchas PCs, PDAs and oth er mo bile devices.

Internet f lntranet

MlM middleware

The network transports data over a local area network such as Ethernet or a wide area (e.g.GSM/GPRS) between M 2M h ardware and the central managem ent application.

This is a TCP/IP based infrastructure for g lobally sharing data from remote devices with oneor m ore ce ntral applications.

Middleware provides routing and buffering of data between remote m achines and centralITsystems. Email could be con sidered as a simple form of middlewa re, whic h m anages thedata exchange between dissimilar systems. M2 M specific middleware offers a higher leve l ofcomm unication man agem ent with interfaces for different types of machines andIT systems.

M2M services These specialised services for M 2M applications offer automa ted delivery of information,hosting, access to information, autom ated reports and alarm notification.

witho ut useof W NechnologyAnd we have t o r eme mber t h a t IP addresses are

dynamically assignedby ISPs and GPRS operators, wh ichcan hange from connection o connection.

In ternet and/or GPRS based M2M solutions favour themmote MZM hardware to act as a client. A s a resul t cent ralsite polling s not deally suited tothis environment.

Wi than nternet-based solutionin place, it is mpo r t an tto r e a l is e t ha t t he In t e rne t i s no t a c loud of i n f i n i t e

bandwidth - as i s depicted in many diagrams, where anydata poured in to it w il l magica lly appear at the application!

Contention needs to be managed across the end-to-endsystem as a whole.What needs to be considered is:

Network access i s here suff ic ient s ignal s t rength toestablish a GSWGPRS connection?l n t e r n e t ~ - E a c h P o P h a s a ~ t e m o d e m l i m i t asedo n a r a t i o to P C u s e r s w i t h r a n d o m c o n n ec t io nschedules. Wh at happens if thousands of m achinesconnect on a default schedule, for exam ple, at midnii t?

Network - T h e I n t e r n e t i s m a d e np o f m a n yin t e rconnec t ed ne tworks w i t h ow bandwid th po in t ssuch as he leased ine connection o he central server,par t icular lyif this is shared w i t hPC user traffic.Seruer access

~ TCP/ IP and t he In t e rne t supportssimultaneous communication at the server applicationove rcoming t r ad i t i ona l po l l i ngbot t lenecks . Can theserver appl ica t ion sca le to many thousands ofs imul taneous T CP connect ions across one or m o r e

servers? +

The problem withthe existinginfrastructureis that it was designedfor humanswho can make simpledecisions to recover from adverseconditions 9 9

IE E Computing 8 Control Engineering 1OaokrlNwember 2024 I 3

8/2/2019 01394737

5/7

I A

- -I I

fi g 4: Embedded mach ine designers generally don't understand central IT integration while IT depaftmenb and/SF%generally don't u nderstand embedded machines (especially non-PC based). This creates a potential 'void

Without system level consideration, there is a potential'void' between remote machines and the central lT softwaresystems, where embedded machine designem do not largelyunderstand ce ntral IT integration. At the same time, ITdepartments and ISPs do not understand the needs ofembedded machines, especially where they ar e non-PC

based. The infras tructure between them was designed forhuman users where simple decision making underpins theresilience.

As the human element is removed from an M2Msystem, the decision-making for maintaining theresilience needs to be added back. How can this be buil tin and automated with out knowledge of t he end-to-endsystem as a whole?

It is quite easy to trial M2M demonstration systems;however deploying resilient solutions in large scale volumeis non-trivial. Machines are expected to operate in the fieldfor many years without human intervention.

By adding M2M connectivi tx the system design of themachine migrates beyond the physical machine itself to thecentral application and the infrastructure between.Communication management therefore needs to beconsidered at t he end-to-end system level and integra tedwithin each element.

A STRATEGIC APPROACHLet's us put the technical issues to one side for the moment- these are not important until the new 'connected'products and services have been defined a nd specified.Therein lies the problem ~ with the potential forsignificant change towards service, companies often do not

know what they want.hif2M projects many grand visions of machines talkingautonomously with enterp rise systems - machines thatdetect failures and notify cent ral applications, whichautonomously check inventon, order new parts, schedulemaintenance visits and debit customer accounts where therepair is made, often before the customer knows of theproblem!

These are very real, angible solutions, but they may bebeyond the first steps that a company interested inimplementing M2M should take. The solutions m complexand involve a mixture of business and technology

Complex solutions, where both the technology and thebusiness case are not fully understood, can lead to projectfail=. For all of the tangible visions that M2M technologypaints, urrent reality shows that M2M projects often fail tosucceed if a Return on Investment (ROI) cannot bedemonstrated in the short-term. Therefoq those who wish

Moption Phne Volume Deployment

rime to market minimal Investment & risk keyh e r e PROVIDING a burloerr model

Higher integration. lower wI t and optimumIundonaliQ key h e n usiness model PROVEN

fig 5: The blueprint far success of M2M implementation

I 4 IEE Computing 8 Control Engineering Ocfober/November 2004

8/2/2019 01394737

6/7

M2M

to implement an M2 M strategy should be s ure they canidentify a business case that shows a return on investment,and that they understand the technology issues.

The solution is obvious but offen ignored - keep it simple!Companies should take are not to get carried away with thegrand vision, but instead focus onfmdinga simple solutionto a specific core problem, and implement the solution usingoEtheshelf components so that the solution can be rialledwith fast time to market.

This enables the product roadmap to be specified,designed and deployed in volume. The emphasis here is toreduce cost through design integration and optimisationover time, when the business case and requirements arevalidated. With solid foundations, the project is not a t riskduring the longer pbax of rollout development. S i p l e stepsthat enhance tbe business are key; rather than wholesalechanges. Fig 5 summaries the blueprint for success.

In summary, this highlights two phases of M2Mimplementation where the fundamental drivers andtechnical solution differ. The first is adoption, wherecompanies are eeking to prove a business model that canyield a healthy le turn on inveslment. The emphasis lis witha quick trial at minimal investment and risk.

This is followed by a second phase, volume clepbyment,where companies have proven their business model,delivered growth returns and need optimised costs. Theemphasis lies with higher integration, volume costreduction and optimum functionality

SOLUTIONSWe will discuss solutions for each element of the end-to-end system and h i g h l i t which are appropriate for theadoption and volume deployment phases.

Adoption suits off-the-shelf box products, which canexternally adapt legacy equipment without requiring *

--End-to-end system

via?he network modem.

Transports logged data over GSM/GPRS networksinto an off-the-shelf web application, whichsupports central database storage, alarming,reporting, notification and administration.

IEE Computing (i Control Engineering OdoberINavember 2 W 4 45

8/2/2019 01394737

7/7

OEM Modules

Design license

Custom design

VOLUME DEPLOYMENT PHASEI

Integrated within the OEM equipment , which maintain flexibility an d fast time to market whileminimisine the risk associated with wireless performance and R & Tr E regulatory conformance.

M 2 Msolution providers have license proven designs for high volume applications offering leastrisk and minimum cost of implementation.

The alternative is for companies to implement lengthy custom developments by integrating thecore design elements incorporating the GSM/GPRS engine (e.g. Sony Ericsson GM47).associated power supply conditioning, SIM holder, processor, Real Time Operating System

(RTOS), TCP/IP stack and associated data transfer protocols.

hardware changes to the machine. M7.M baniware supportsscalable evels of in tegration, dependant on the intelligencewithin the attached machine (see Off-the-shelf pmducts forthe adoption phase' table, pas).

Higher integration and reduced costs can be achieved atthe volume deployment phase by migrating the box-baseddesign in stages as part of a s truct ured roadmap (see'Volume deployment phase' table, above).

Adoption suit s the use of exist ing platforms andinfrastructure o provide route and bufering functions ona service basis, without the need for large up frontinvestment. These are ypically available from traditionalISPs in the form of email, FTP nd HTTP servers.

Large scale volume deployment benefits from amanaged communication platform to ensure high servicelevels and res ilience, whether obtained under service aspart of a solution providers' offering or purchased outrigbt.

Although traditional SPs and server farms offer hostingservices for IT based solutions, specialised ISPs andApplication Service Provider's (ASPS) are becomingprevalent offering M2M specific services .

The adoption phase suitsapplications hat already existor can be developed quickly with core minimalfund1onality

For example, a simple utility could obtain data ff is h memail, FTP or HTTP servers and populate an existinglegacy application or utility program. Thi s could be assimple as populating an Excel spreadsheet th at enableso m e eports o be generated.

On the other hand, a n end-to-end system could be anoff-the-shelf system comprising a web application thatcommunicates with specialised M2M hardware whichsupports central database storage, alarming, reporting,notification and administration. These are available fromASPS under service or h m 2M solution providers.

Although M2M solution providers and ASPS offerstandard products suited to adoption and volumedeployment, in many cases highly custom solutions, egacysystem or enterprise integration is required to meetcompanies' bespoke needs. These can be developed in-houseor using thM party soilware systems ntegrators. This mayinvolve an alliance comprising the OEM,MZM hardware,middleware, wireless operator and software solutionpartners to deliver he complete system.

Open standards are clearly important for M2Mproliferation, pa rt im hl y with the wide alliance of parinerstha t are potentially needed to deliver a complete system.With such diverse applications under the M2M umbrella,there is no open standard that meets the needs of d. TlYdue to its widespread use, enables automated open dataexchanges using Web services , SOAP and XML. However,simple solutions based on SMS, email, ETP an d H W anbe deployed to bring benefits in a shorter amount of time.

Successful adoption of M2M technology require s astrategic approach to ensure that the technical solution isbalanced with the business case to demonstrate an earlyreturn on invesbnent. Simple solutionswork best where thetechnology can be proven quickly and the benefits easilyunderstood. This creates time to reduce cost and optimisepmducts without risk, enabling new business opportunitiesto be created.

Camtech Holdings's pWEAVE application can e used orrapid M2M technology adoption. It enables companies totri al machines rapidly by collecting data, distribu tingco nf ii at io n, processing alarms, automating SMS/emailnotificaions, generating online reports, xporting data andmanaging the system as a whole.

Steve Whitehead is Technical Director, Comtech HoldingsLtd. Wvmcomtech.uk.com

46 IEE Computing E Control EngineeringIOctnberlNouember2004

http://wvmcomtech.uk.com/http://wvmcomtech.uk.com/